ACOUSTICS AND GRAVITATION. 



105 



embodied in the mean; but on the hot days the difference is enormous. One 

 may note that the longest excursion exceeds 7 scale-parts ; the smallest is below 

 2 scale-parts. 



To compare these with the interior temperatures of the laboratory would be 

 without avail, for the thermostat data do not vary as much as i per day, as a 

 rule. Moreover, these air-temperatures are mean data with no direct bearing, 

 whereas the weight M reciprocates directly, in radiation, with the walls of the 

 room. Hence the external or atmospheric mean temperatures are pertinent. 

 They were kindly furnished by Charles S. Wood, meteorologist of the U. S. 

 Weather Bureau Station in Providence, Rhode Island. These mean temper- 

 atures for the day are inserted in the top of figure 121, in degrees Fahrenheit. 

 The numbers attached show the rainfall in inches. A scale twice the one 

 taken would have facilitated comparison. 



Mere inspection of the two curves for July shows them to be identical in 

 character of variation, except that the lower curve follow? the upper with a lag 

 of one to three days. In the region a there are two maxima differently 

 developed; also in the region b. The cooling effect of high precipitation is 

 marked in both. All this is in some respects what would be expected, as it 

 takes any external isotherm a day or more to penetrate the 30 inches of wall 

 thickness in order to enter into reciprocal radiation exchange with the ball M. 



It is a little surprising, however, that the high temperatures are distinguished 

 from the low temperatures in figure 121 by such large a.m. and p.m. dif- 



